In Hancox v Sutherland and others [2026] EAT 139, the Employment Appeal Tribunal (“EAT”) provided some much-needed guidance on the use of generative AI by parties, particularly litigants in person (“LiPs”), when preparing documents for submission to courts and tribunals.

Given the prevalent use of generative AI in employment litigation, the guidance sets clear expectations on the responsibilities of LiPs and the consequences should these responsibilities fail to be complied with. 

Use of generative AI in employment tribunal litigation – what is the current picture?

Free to access, generative AI tools such as ChatGPT are regularly relied on by LiPs who bring employment tribunal (“ET”) litigation. Responsible use of generative AI can facilitate access to justice – providing LiPs who do not have access to professional legal advice with support when drafting legal submissions and reviewing documents to be submitted to the tribunal. However, there are significant risks of irresponsible and unchallenged use of generative AI, as the output of generative AI can often be: 

• Unfocussed: resulting in long, complex and often irrelevant submissions which do not focus on the key issues in the case. 
• Misleading: with documents and applications submitted to the tribunal not reflecting the key factual, evidential or legal position in the case due to deficient prompting.
• Affected by biases: as generative AI strives to provide a ‘positive’ outcome for users, the relative strength of positions or prospects of success are often over-exaggerated. 
• Non-compliant with procedural requirements: such as case management orders, ET procedural rules and EAT Practice Directions. 
• False or distorted: and contain ‘hallucinations’ (e.g. made-up case citations or incorrect statements of the legal principles established by cases). 

These risks are currently being borne out in the ET system. The prevalence of AI use by LiPs has resulted in lengthy claim forms in which a ‘kitchen sink’ approach is taken, with multiple causes of action pled. Additionally, more complex claims with no qualifying service requirement (such as whistleblowing and discrimination) are being pursued, which makes defending claims a more time and cost intensive exercise. In particular: 

• The latest ET quarterly statistics (covering the period from April to June 2026) note that single ET claim receipts have increased by 28% from the previous quarter, which is in part due to the availability and use of AI to prepare claims. 
• Additionally, there has been an increase in ‘complex’ open-track claims, such as whistleblowing and discrimination, which employment judges have noted is likely due to the use of AI by LiPs.
• Applications for interim relief, an emergency remedy in a limited category of automatic unfair dismissal claims, have sky-rocketed. While ETs across Great Britain would typically receive 20 applications a year, they are now receiving 20 a month. This has resulted in the President of the Employment Tribunal issuing guidance on applications for interim relief, that cites the significant increase in applications which “often indicat[e] the use of AI”. 
• Guidance for judicial office holders on AI has been produced which notes that, where appropriate, judicial office holders should inquire whether the LiP has undertaken checks for accuracy on AI generated content, and that LiPs should be informed that they are responsible for what they put before the tribunal or court. 

Prior to the judgment in Hancox, much of the case law on parties’ use of AI was in the context of regulated lawyers’ use of AI and their professional and regulatory duties. While LiPs are not bound by the same professional and regulatory duties, they are required to assist the Tribunal to further the overriding objective which includes “dealing with cases in ways which are proportionate to the complexity and importance of the issues”. However, there are no specific ET rules or guidance regarding LiPs’ use of AI.   

Background 

The facts of this case are somewhat unremarkable, but will likely be familiar to those responding to claims in which a LiP has used AI to produce documents submitted to a tribunal. 

• Mr Hancox (“H”) brought claims of automatic unfair dismissal, unfair dismissal, whistleblowing detriment and disability discrimination against four individual respondents. 
• Following an unsuccessful application for interim relief, H embarked on a LinkedIn campaign criticising the four respondents and publicising information relating to his ET claims, including disclosed documents. Consequently, H’s claim was struck out on the basis that his conduct was scandalous, unreasonable and vexatious and a fair trial was no longer possible.
• H appealed this decision and the matter was listed for a preliminary hearing before the EAT. Prior to the hearing, H submitted a 300-page skeleton argument created by ChatGPT, which was non-compliant with the EAT’s Practice Direction. H in fact chose not to rely on the skeleton argument at the preliminary hearing, accepting that he had not checked the document for accuracy, and that it lacked credibility. With the assistance of his barrister, H advanced submissions on one ground of appeal only. 
• Despite confirming at the hearing that he did not wish to pursue further grounds of appeal, H sought to pursue a further ground of appeal. The EAT also noted that both before and after the preliminary hearing, H made numerous (often repetitive) applications and engaged in excessive correspondence with the EAT.

The EAT’s decision 

While the EAT held that H’s one ground of appeal advanced at the hearing was arguable and would proceed to a full hearing (H’s application to rely on an additional ground of appeal was dismissed), the headline of the EAT’s judgment was the guidance for litigants using AI. 

After a review of the relevant case law and guidance on AI use, Deputy Judge Pilgerstorfer set out guidance providing that litigants who use AI should ensure as a minimum, that all documents submitted must: 

1. comply with applicable procedural rules; 
2. have been checked as thoroughly as the litigant or representative is reasonably able for accuracy, ensuring that factual, evidential and legal points (including references to authorities) are correct and that the tribunal is not misled; and 
3. contain only relevant points, with the focus being on the central or best arguments, presented in an easily comprehensible manner and avoiding undue repetition. 

The EAT further noted that there is no principled reason why a LiP should not take reasonable steps to use AI responsibly, and that while a LiP will not usually be subject to the professional duties of regulated lawyers, the underlying principles of personal responsibility and accuracy apply to all persons who submit documents to a court or tribunal.  

The EAT also clarified that submitting generative AI output without checking it, or placing the checking onus on another party, is not acceptable. Where there are concerns that generative AI output has not been checked (or checked inadequately), judicial enquiries and potential sanctions are likely to ensue.

So what for employers? 

It is hoped that the guidance will be an effective ‘setting of the goalposts’ with respect to LiPs’ use of AI to prepare documents submitted to tribunals. The EAT’s position on parties’ (including LiPs’) responsibilities is clear – and where these responsibilities are not complied with, parties will be at risk of tribunals placing no (or limited) weight on AI generated content and possibly further sanctions, including the strike out of proceedings, the making of unless orders, and/or costs orders. Employers who suspect that LiPs are submitting AI generated content to the tribunal without undertaking the necessary checks should: 

• keep a clear audit trail of the suspected AI generated content; 
• challenge the LiP on whether the content is AI generated (with a particular focus on unclear or irrelevant allegations, lengthy correspondence, undue repetition and hallucinated cases) and, if so, whether they have checked this content, highlighting: (i) the LiP’s personal responsibility for the content of documents submitted to the ET; and (ii) the guidance provided by the EAT in Hancox; and 
• where (despite the challenge raised) the LiP continues to submit unchecked, unfocussed AI generated content to the Tribunal, raise this with the Tribunal. 

If this is something you would like to discuss further, we would be delighted to hear from you. Please do not hesitate to get in touch with our employment law team.